Internet research does not seem to be a new topic, since it is relevant from the beginning of the spread of the internet for everyday use in the early 90s. However, apart from great progress in the design and implenmentation of effective and efficient seach engines, there are challenging issues which are still far from being solved and which even gain importance due to the rising omnipresence of the internet.
One such aspect is the assessment of the quality of information and their sources. And this issue continuously changes its appearance due to the change in the ways how information is presented, hosted, maintained, distributed, and accessed. E.g., the rise of social media turned every internet user also into a publisher, which has obvious impact on the quality and reliability of the presented information. Fake news is just one extremely ugly face of this phenomenon. Since each day more and more aspects of our lives rely on the internet, this problem further intrudes our everyday life. In particular, already today our working life is highly affected by it.
The abundance of content in the internet is great, but it is also a problem: When researching, the essentials have to be filtered out of a mass of data. This requires the ability to distinguish relevant from irrelevant, balanced from biased, objective from subjective information, and serious from dubious sources. This represents an essential part of the information literacy which students of the Lucerne University of Sciences and Arts learn in modules such as “Targeted learning in the company”, where they are asked to research on certain given or self-selected topics and, based on this, to create a plan of the learning content. Lecturers are there to support them as coaches. In order to be able to judge the search strategy and to provide helpful feedback, the coaches need detailed information about the research behavior of the students. However, since they cannot constantly monitor the working behavior of the students, this information is mostly lacking. Therefore, researchers of the university set out a research project, whose long-term goal is to develop a tool that logs, processes, analyzes, and visualizes the research behavior of a user. The purpose of this tool is to assess whether the search was targeted and efficient and whether the gained information is objective, accurate, complete, and unbiased.
Among others, this research raises the question how the quality of a website can be judged. Here, quality means objectivity, accuracy, completeness, trustworthiness, and unbiasedness of the information provided by the site. Since the judgement of these criteria is difficult and time consuming, it is unrealistic to rate a significant number of websites manually. Therefore, the idea is to define a statistical measure which gives at least a rough estimate of the quality of a website in the above sense. One approach may be to manually evaluate a certain number of websites, and then follow the links to pass the quality measure on to other websites. The idea is that high quality websites are more likely link to other high quality websites. There have been already attempts to define such measures such as MozTrust. The MozTrust of any website can be determined using checkmoz.com. However, this and similar measures have only be used as a tool for SEO so far. As a consequence, it is not so clear whether they measure true quality or merely search engine ranking.
For this reason, a particular focus of this research project is to improve and refine these measures to more accurately rate a website with respect to its information quality. For this endeavour, the researchers look for an industrial research partner, who is willing to invest a significant amount of effort into this goal, and on the other hand will be rewarded with a lot of attention for its contribution to such a major achievement for science, economy, and society as a whole.
